Improved Segmentation of the Intracranial and Ventricular Volumes in Populations with Cerebrovascular Lesions and Atrophy Using 3D CNNs

February 2021
Journal: Neuroinformatics
Lead Author: Emmanuel E. Ntiri

All Authors: Emmanuel E. Ntiri; Melissa F. Holmes; Parisa M. Forooshani; Joel Ramirez; Fuqiang Gao; Miracle Ozzoude; Sabrina Adamo; Christopher J. M. Scott; Dar Dowlatshahi; Jane M. Lawrence-Dewar; Donna Kwan; Anthony E. Lang; Sean Symons; Robert Bartha; Stephen Strother; Jean-Claude Tardif; Mario Masellis; Richard H. Swartz; Alan Moody; Sandra E. Black; Maged Goubran

Successful segmentation of different parts of brain images is of critical importance when studying neurodegeneration through neuroimaging. We compared several robust algorithms that use their own unique methodologies to achieve this segmentation. Our own models were tested on a large dataset of older adults with varying degrees of cerebrovascular lesions and degeneration. Our resulting multi-contrast models outperformed other methods across many of the evaluation metrics.